8.851 Homeworks 8 & 9

advertisement
8.851 Homeworks 8 & 9
Iain Stewart, May 6, 2006
This will be the last problem set. It includes more problems than usual
and will count as two problem sets.
Problem 1) Non-Relativistic QCD at the ILC
√
Consider e+ e− → tt¯ near the threshold for tt¯ production ( s ∼ 2mt ∼
350 GeV) at the International Linear Collider. In this region
√ the top quarks
are non-relativistic with small residual energy, with 2E = s − 2m ∼ mv 2 ∼
5 GeV, and non-relativistic momenta p~ ∼ 30 GeV. The top quarks have a
width for t → bW that is numerically similar to mv 2 , so we will include their
total width as part of our LO potential Lagrangian:
Lp =
�
ψp†
p
iΓ
p2
+
+ . . . ψp + (ψ → χ) + . . .
i∂ −
2m
2
�
�
0
(1)
where the . . . indicate higher order relativistic corrections as well as the
potential interactions. The Coulomb interaction is relevant so we will include
it too. The cross section can be obtained from the imaginary part of a Green’s
function computed from this action, and in the MS scheme is
m2
G (a, v) =
4π
0
�
−i v
1
ia
− + ln 2 + γE + ψ 1−
i v − a ln
ν
2
2v
�
�
�
�
���
.
(2)
√
Here a = CF αs (mν), v 2 ≡ ( s − 2m + iΓ)/m, m = mtop . The polylog function ψ(x) = d/dx ln[Γ(x)]. I have written the renormalization scale µ = mν
in terms of a velocity renormalization parameter ν.
a) Lets see which diagrams generate G0 (a, v). Consider the forward scattering diagrams from the time-ordered product of a production and annihilation current, T Ji† (0)Jj (x). For e+ e− → tt̄ through a virtual photon we
have Ji† = ψp† σi (iσ2 )χ∗−p , but feel free to ignore the spin for your calculation.
Compute the one-loop diagram and show that it reproduces the iv term in
Eq. (2). Of all the diagrams that generate G0 (a, v) which have UV divergences? Which diagram gives the term with a logarithm?
b) Using mathematica
plot the normalized cross section R = 8πImG0 /m2
√
as a function of s. Taking αs (mZ ) = 0.118, m = 175 GeV, renormalization
velocity ν = 0.15, and using one-loop running for αs (mν):
i) Plot R for the values Γ = 2 GeV, Γ = 0.7 GeV, Γ = √
0.1 GeV, Γ =
0.05
√ GeV. Describe the physics behind what you observe for s > 2m and
s < 2m.
√
ii) Take Γ = 0.05 GeV. Analyically determine the location in s where
you expect the first and second peaks (without using Eq. (2)) and see if it
agrees with your plot of R for this Γ. (Hint: Remember that we’re doing
QCD with a Coulomb interaction. In QCD the coefficient of the Coulomb
potential in the color singlet channel is Vc = −4παs (mν)CF with CF = 4/3.)
iii) Taking the realistic value Γ = 1.5 GeV explain how this cross section
could be used to measure the top mass m.
Problem 2) SCET Operators with Collinear Quarks
a) Start with the QCD Lagrangian for a massive quark and decompose D
/
in terms of n, n,
¯ and ⊥ components. As in lecture, write ψ = ξn + ξn¯ and
determine which terms are non-zero. Keeping all the terms integrate out the
field ξn̄ to generate an effective action for ξn .
[With power counting m ∼ p⊥ ∼ Qλ this is the starting point to derive the
action for a massive collinear quark, ie. prior to decomposing the gluon field
into collinear and usoft pieces and prior to distinguishing between large and
small momenta. The steps here are similar to those outlined in lecture except
that you are keeping the mass.]
b) Consider the current for a b → u transition. In QCD J = uΓb.
¯
For SCET
we did a matching calculation to find the leading order current
J (0) = ξ¯n W Γhv ,
(3)
where W included terms involving the order λ0 collinear gluon field n̄ · An . In
lecture we explicitly computed the term in W with one n̄ · An field and wrote
down the result for any number of n̄ · An fields. Do the matching computation
for two n̄ · An fields (by expanding QCD diagrams with offshell propagators).
Bonus: Verify that the result for one and two n̄ · An fields agree with the
momentum space Feynman rules for the position space Wilson line
+
�
W (y ) = P exp ig
where P is path-ordering.
�
0
�
ds n
¯ · An (s¯
n + y) ,
−∞
Problem 3) Decoupling of Ultrasoft gluons in SCETI
Consider a Wilson line built from usoft gluons
�
Yn (x) = P exp ig
�
0
µ
µ
�
ds n · Aus(sn + x ) .
−∞
(4)
It satisfies Y
n† Yn = 1 and has the equation of motion n · Dus Yn = 0. Start
with the leading order Lagrangian, L(0) for a collinear quark in SCETI . This
action gives eikonal couplings to ultrasoft gluons. Make the field redefinitions
†
(0) (0)
(0)
ξn = Yn ξn(0) and An = Yn A(0)
n Yn to obtain a Lagrangian L (ξn , An ). Show
explicitly that this new Lagrangian has no coupling to n · Aus gluons.
Problem 4) Decoupling of Soft and Collinear Gluons in SCETII
Consider an operator with one collinear quark and one soft quark in SCETII .
This case differs from SCETI in that soft gluons knock collinear quarks offshell and collinear gluons knock soft quarks offshell. To be definite lets make
the soft quark a heavy quark from HQET, hv , and the energetic quark a
massless collinear quark, ξn . The SCET operator will be
(ξ¯n Wn̄ )Γ(Sn† hv )
(5)
where W is the Wilson line we saw in problem 2b, and Sn is an exact analog
of Eq. (4) but with As gluons rather than usoft gluons. To find Eq. (5) requires tree level matching (to determine the direction and gluon components
appearing in the Wilson lines), and gauge invariance (to ensure that loops
do not spoil the structure so obtained). The first non-trivial term in perturbation theory have one soft and one collinear gluon. By computing the
graphs in Figure 1 expanded to LO, verify that the three-gluon interactions
are responsible for putting the gluons in the right order in Eq. (5).
(a)
(c)
(b)
qs
qc
µ,a
µ,a
µ,a
qc
ν,b
qs
qs
ν,b
qc
ν,b
Figure 1: QCD graphs with collinear, qc , and soft, qs , momenta.
Problem 5) Pion Light-Cone Distribution Function
Factorization of degrees of freedom allows us to describe processes in terms
of simpler (universal) objects. One such object is the pion light-cone distribution function. In QCD we can define
b
µ 5 τ
¯
hπ (p)| ψ(y)γ
γ √ Y (y, x) ψ(x) |0i = −ifπ δ ab pµ
2
+...,
a
�
0
1
dz ei[zp·y+(1−z)p·x] φπ (z)
(6)
Here fπ ≃ 131 MeV and the field ψ denotes the isospin doublet (u, d). Connµ ,
sider the path from y µ to xµ to be light-like, (y − x)2 = 0, with y µ = y¯
µ
µ
x = xn̄ . Y (y, x) is a Wilson line along this path. For the leading order
collinear operator in SCET
τb
a ¯
hπn,p
|ξn,p1 W n/̄γ5 √ δ(ω−P¯+ ) W † ξn,p2 |0i
2
= −ifπ n·p
¯ δ ab
�
0
1
n ] φπ (z) ,
dz δ[ω−(2z−1)¯·p
(7)
where P¯+ = P¯ † + P̄. Eq. (7) can be obtained from Eq. (6) by projecting
onto the LO term with collinear quarks and taking a Fourier transform with
x = −y.
a) Using the power counting for the collinear fields in λ = ΛQCD /Q, together
with mass dimensions, count the powers of ΛQCD and Q on the LHS and
RHS of Eq. (7) and verify that φπ (z) is O(λ0 ) and is dimensionless.
b) Under charge conjugation, C −1 ξn,p (x)C = −[ξ¯n,−p (x)C]T where C is the
usual charge conjugation matrix. For a π 0 state use charge conjugation together with Eq. (7) to prove that φπ0 (z) = φπ0 (1 − z). Physically what does
this mean? What do you have to assume to prove that this is true for the
π + and π − ?
c) Bonus: Carry out the steps to go from Eq. (6) to Eq. (7) ignoring the
ellipses.
Download